FlexiReID: Adaptive Mixture of Expert for Multi-Modal Person Re-Identification

Multimodal person re-identification (Re-ID) aims to match pedestrian images across different modalities. However, most existing methods focus on limited cross-modal settings and fail to support arbitrary query-retrieval combinations, hindering practical deployment. We propose FlexiReID, a flexible framework that supports seven retrieval modes across four modalities: RGB, infrared, sketches, and text. FlexiReID introduces an adaptive mixture-of-experts (MoE) mechanism to dynamically integrate diverse modality features and a cross-modal query fusion module to enhance multimodal feature extraction. To facilitate comprehensive evaluation, we construct CIRS-PEDES, a unified dataset extending four popular Re-ID datasets to include all four modalities. Extensive experiments demonstrate that FlexiReID achieves state-of-the-art performance and offers strong generalization in complex scenarios.

Identifying people across different types of images—such as color photos, infrared images, sketches, or even text descriptions—is a difficult task, especially in real-world settings where information may be incomplete. This paper introduces a new system called FlexiReID, which can flexibly search for a person using any mix of these information sources. For example, if only a sketch and a short description are available, the system can still find the matching photo. The key idea is to combine expertise from different specialized models and adapt based on what information is available. To support this system, the authors also created a new dataset that includes four types of data. Experiments show that FlexiReID works better than previous systems, especially in complex or low-quality situations. This research makes person search more practical and reliable in real-world applications like public safety and smart cities.